MSA TT P-8357
6-0 c/<e. ^ ol
ELECTROMETER TUBES: PART II
by H, Dolezalek
BECAUSE OP COPYRIGHT RESTRICTION THIS TRANSLATION HAS NOT BEEN
PUBLISHED. THIS COPY IS FOR INTERNAL USE OP NASA PERSONNEL AND
ANY REFERENCE TO THIS PAPER MUST BE TO THE ORIGINAL GERMAN SOURCE
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
WASHINGTON February 1963
/Published in Archiv fuer Technisches Messen ,
Ho. J, 8334-4? January 1962, pages 19-22/
EISCTHOMEIEB TUBES s PIET II
Hans Dolezalek, Diploma Pbysicisi;
Meteorological Observatory of the German ?feather Services, Aachen
The report begun in Part I (ATM Ho, J-8334-3j December I96I) is
continued with the description of systems with electrometer tubes.
Yery many designs have been developed for the arrangement of
amplifiers for electrometer tubes of which we shall discuss only tvro
In all cases concerning continuous measurements with electrometer
tubes which is not possible without a certain degree of automation,
utilization of a bridge circuit is the preferred means (5-g)» However,
if a short-interval measure is visually controlled by the observer who
can manually regulate input voltages, a straight-line circuit is then
preferable because of consuming less input current and therefore more
g) - Bridge Amplification for Electrometer Tubes . Utilization of a
l?ilheatstone bridge with two electron tubes and two working resistances
for direct-current amplification was apparently first proposed by Wold -
/16_/, This was further developed subseq_uently, including designs for
bridge circuits with only one electron tube and replacing the others by
greatly varying types of circuits. After electrometer tubes began to be
available, they also were employed for such bridge amplifiers. The
advantages of a bridge amplifier arej
I. Disturbing influences of input voltage variations are reduced
or canceled out by tubes sufficiently similar to each other within the
limits of accuracy so given. If the two tubes are someTirhat different,
which is the general rule, similarity can be improved by balancing methods.
If the circuit is correctly designed, zero drift due to ageing and tempera-
ture variations is reduced.
II. In raaror cases, a bridge circuit permits compensation of
disturbances which become mixed with the impulse already through the
generator or circuit. This is successful if a reactive generator (with
reactive line) can be employed parallel with the measurement generator
(with measurement line). This retains advantage I.
III. If the impulse, is symmetrical to ground and the two components
of tension to ground are not of interest, the bridge amplifier can be
used as push-pull amplifier (impulse voltage between the control grids
of the two tubes). The summation value then appears in the output. This
also retains advantage I.
A standard wiring diagram of a bridge amplifier is shown in Pig. 2
/sic-?~should be Fig. 1 — Tr,J. In principle, both electrometer tubes
should always be provided with equally high ohmic resistances at the
control grid — and equally large capacities — because only then is it
possible to achieve the optimiim of bridge balancing and only then do
advantages II and III become at all perceptible. As indicator or recorder,
either a high-ohmic voltmeter or a high-ohmic tension recorder is used
(cf . Pig. 2) which compares the two anode voltages, or the two working
resistances are replaced by the two coils of a differential galvanometer.
Further direct-current amplifier stages which are to be connected
may, in turn, be designed as bridge amplifiers.
Proper operation of the bridge depends on the similarity of the
two electrometer tubes utilized. Unfortunate ly^ the manufacturers of
electrometer tubes will not make the effort of selecting pairs of tubes
as similar to each other as possible and leave this to the consumer.
Utilization of a twin-tube v/^ith a common cathode is obviously the desirable
arrangement (cf. 6-d}.
Even selected pairs still have differences which are obviously
greatest in regard to filament emission, lie are therefore unable to get
along without so-called "balancing," This consists in varying the filament
currents in both tubes in the opposite direction by "systematic trial and
error" until no further change in. the differences of anode voltage occurs
when the common filament .voltage is changed (cf . Schintlmeister flij for
Fig, 1, Standard wiring diagram of the bridge amplifier.
Such balancing is obviously possible accurately only for a single
operating point. For this, we select about half of the average value of
the magnitude to be measured subsequently. Balancing similar to that of
the filament currents can be effected also for the anodes by varying the
Value of the working resistances also in the opposite direction (potentio-
meter connected at the upper branch of the bridge). Anode balancing
detunes the filament balancing made earlier which must then be repeated
again until both are satisfactory but filament balancii:^ is generally-
sufficient. If the tubes have space-charge grids, these can be advanta-
geously used for balancing (cf. 6-a).
^) Straight Amplification for Electrometer Tubes . If we are forced
to economize on battery weight which results in reduction of the number
of tubes, then the straight amplifier is preferable. This eliminates argr
compensation of input-voltage variations and all other advantages of the
bridge amplifier so that the straight amplifier should be used in
principle only when all voltages are continuously controlled and can be
adjusted manually. In certain cases, it may be used when direct observa-
tion is not possible, provided that all input voltages are recorded
sufficiently accurately so that it becomes subsequently possible to decide,
depending on the necessary accuracy of measurement which parts of the
recorded data can be utilized. The arrangement should be such that it
prevents any regeneration effects as far as possible. Fig. 3 /sio — should
be Fig. 2 — Tr._/ shom^s the wiring diagram of a straight amplifier which
needs no further explanation.
Fig, 2, Standard wiring diagram of the straight amplifier.
Here it will be best to use triodes, not only because this saves
space-charge grid current and/or grid-screen current but because they
furnish a higher plate current and can therefor© be used more easily
without additional amplification. A measuring instrument and a regulating
potentiometer should always be coupled in the circuit,
i) Amplifier with "Floating Grid ." Ever since electrometer tubes
have become available, the "floating grid" method has teen proposed again
and again. It consists in entirely eliminating a defined grid-leakage
resistance. In order to properly evaluate the method, we must go back
to the composition of the ccmponents of the grid current.
In the operation of electrometer tubes, several extraneous resist-
ances are inevitably parallel (cf . 4-c) to the "tube-inherent" insulation
resistance (group 1 in 4-a). They all act jointly and cannot be sepa~
rated. If the generator .does not produce any tension or produces a
constant tension, then the potential of the control grid also remains
constant as long as the grid current does not change. In other words,
when tube and parallel resistances are coupled, the grid current is zero.
In principle, this is possible for any grid potential but is not an
indication that the grid potential adjusts itself to the zero point of
the characteristic of the "total grid current" (Fig. 1). On the contrary,
the grid potential corresponds to the grid bias. This does not change
fundamentally if we now make the defined grid-leakage resistance larger
and larger. In principle, it should be smaller than the resistance of
the electrometer tubes and of the generator and line leakages, etc.,
because it is the only one defined and (relatively) accurately known.
We sacrifice this advantage of a relatively known leakage resistance if
we assume E„ as infinite.
From this point of view, we can say that the floatijag grid arrangement
does not exist in principle* f/henever an approach toward it is made, we
only attempt what should always be done when using electrometer tubes for
maxim^^m efficiency which is to derive the maximum value of the still
permissible grid-leakage resistance through estimating the tube-inherent
insulation resistance and yet to rest so low with it that the unknown
variations of the tube-inherent insulation resistance will not cause
The difference in character of the grid-current grou-p 1 (4-h) as
against groups 2 and 3 Jaay already be seen from the fact that groups 2
and 3 are the product of generators for which a supply of power is
necessary and available. The latter is the temperature of the filament
in group 2, and either the liberation of latent energy or of extraneous
radiation in group 3* These characteristics are therefore also not
linear. Only one resistance should be entered in the equivalent circuit
for group 1.
A criterion for the possible use of the "free grid" method for an
electrometer tube can be obtained by comparison of the sum of the grid-
current components 2 and 3 with the grid-current component 1. If 1 is
relatively large compared to 2 and 3, the tube-inherent insulation
distance simply replaces the defined grid-leakage resistance. However,
if one is small, the addition of a defined S„ becomes necessaiy because
the sum of 2 plus 3 varies greatly and the effect of these variations
must be eliminated by a smaller grid-leakage resistance.
From the beginning, authors have warned against the method of the
floating grid. Kleen and Graff under /8/ point out that inevitable
changes of the Yolta effect in the tube uncontrollably displace the
operating point in the course of time and that the trace of the grid-
current/grid-voltage characteristic (in its temporal variability) must
be known because the resistance is not linear. Morton [Wj finds that
the sensitivity of tension of the electrometer tube is less with the
method of the floating grid. Easmussen [\2j believes the method to be
erroneous in principle because the effective grid resistance is then not
constant and only minor amplification occurs in this regard.
Systems containing floating grids are indicated especially frequently
when "inverted triodes" are utilized (of. 6-a).
k) The "Mekapion" Principle . Of the mary circuit designs based on
electrometer tubes, we want to briefly mention only the Mekapion arrange-
ment .,/l47. It does happen that stray effects in electrometer tubes
create an unintended Mekapion effect v/hich is difficult to identify in
searching for the disturbance ^personal communication from G. Hies,
6. Techniques, Designs, Models, Demands
^) Different Electrometer systems and their Manner of Operation .
o() Triodes with Standard Control Grid . The first electrometer
tubes were triodes of customaary design, except for more highly insulated
control grids. Electrometer triodes are still being offered today but
have a higher grid current in general than other systems which will be
A) Inverted Triodes . In principle, the anode voltage of a triode
can also be connected to the grid coil and be controlled by the anode
plate J the "penetration factor" of the control plate influences the
electron flow in the space between cathode and anode grid.
It is to be expected, in this arrangement, that the component 312
(4-a) of the grid current is kept away from the control electrode and the
"grid current" therefore becomes less, f/e can summarize the advantages
of "the inverted triode as follows (Frommhold [\J)x at the same mutual
conductance (as in a tetrode, cf. below), there results a smaller control-
electrode current and the same control-electrode current v/ill produce
higher mutual conductance.
^) The "Plation»" The so-called two-plate tube ("plation") is
based on a principle similar to the inverted triode s the filament is
located between two plane parallel plates, one of which operates as
anode and the other as control electrode.
o) Tetrodes have a positive grid between cathode and control grid.
This "space-charge grid" corresponds 'simultaneously to several purposes.
Because it attracts electrons out of the negative space-charge cloud
ahead of the cathode, it, increases the negative minimtun potential and
therefore affects the effective potential in the space between the wires
of the control grid, makes the latter more positive and so increases the
mutual conductance of the tube. The action of the control grid is now
somev/hat different than in the triode? it also influences the distri-
bution of the current to the two positive electrodes. The space-charge
grid further retains the ions originating on the cathode or in the space
between cathode and space-charge grid v/hich then do not reach the control
grid so that the grid current becomes less. Moreover, the space-charge
grid permits a variety of special systems, e.g., some proposed bridge
amplifiers are designed with only one tube by utilizing the EG. Buhk 
points out that, due to the dropping characteristic of the space-charge
grid current (in some ranges), phase-accurate feedback is possible
already with one tube.
A special advantage of the space-charge grid is the fact that we
can largely compensate the inevitable differences between cathodes in
^) £££i2^£2.' •^^ recent years electrometric pentodes have become
available commercially which obviously permit a relatively high amplifi-
cation of voltage ( /< ^ 30, sometimes indicated as 25O) and work with
very low grid currents, in spite of the absence of a space-charge grid.
They have a specially small plate current,
^^ transistors . During proof reading, we have become aware of an
interesting further development. This concerns a transistor in which the
"grid"- input resistance lies at about 10 ohm. Although it cannot be
utilized directly for electrometric purposes, this would seem to be a
promising development. (C, T, Sah, A new semiconductor tetrode, the surface-
potential controlled transistor. Fairchild Transistor Corp., Palo Alto,
California, I96I, 14 P«)»,
b) Technology of Electrometer Tubes . Requirements for obtaining
high insulation of the control grid consist in the selection of insulation
materials with high specific internal resistance, the treatment of the
surface of these insulation substances and the geometric reduction of
leakage (long and narrow leakage paths). This should be supplemented by
the interposition of grounded metal rings which v^ill prevent the spread
of undesired surface potentials over the insulator although they will not
prevent the flow of charge to ground. Frommhold /47 reports on this in
detail. Best insulation is obtained when the physical realization of the
control grid circuit is as far distant from other feed circuits as possible
(e,g., at the top of the bulb or at the point of the sub-miniature bulb)
and, in some cases, the grid design is surrounded by a glass collar in
order to prolong the leakage path.
For the purpose of reducing other grid-current components, various
other possibilities exist and have been adopted in some cases. We shall
here restrict ourselves to a simple listing because details can be found
in Schintlmeister [VhJ and Frommhold fijt extremely high vaouumj low
cathode temperature, control grids of metal with especially high electron
output efficiency (coating with gold:)| positive electrodes of substances
of low order number | reduction of space angle in which photoelectron
radiation impi3ages on the gridj reduction of space in which extraneous
radiation may have an ionizing effect.
Since electrometer tubes are subject to all the disturbances which
occur in electrometric circuits, positive measures for decreasing material-
electrical effects are necessary. Most important here are the polariza-
tion phenomena in the insulating substances. The nature of these phenomena
which decay r only very slowly ("remanance") is still largely unknown. In
order to decrease their effect, the ratio: of the conductor surface touching
the insulation to that of the surface opposite to the vacuum should be as
small as possible .
These requirements (to be further complemented in 6-d) to some extent
contradict each other and are generally not easy to fulfill. However,
they represent imperatives for eliminating at leas* a part of the "infan-
tile diseases" which restrict the possibilities of utilization of
c) List of Electrometer Tube Models . A listing attempting to be
complete but probably not successful is contained in Table 1, It generally
contains values specified "h-^ the manufacturer which have not always been
checked by us. These values are hard to compare with each other because
they have been obtained under different operating conditions. In particu-
lar, a reduction of the anode voltage below 5 ^ ^^^7 produce appreciably
lower grid currents (at lower mutual conductance) (e.g., T-113 and T-116),
^) Peinands for Improvemen-t of Elec-fcrometer Tubes . In addi-tion to
the demands discussed in 6-13^ the principal demand is for the production
of indiTxdual units of the same tuhe model which are more closely similar.
The obviously desirable goal would be the possibility of replacing
one electrometer tube by another one of the same type without making re-
calibration necessary. This is scarcely possible in highly sensitive
systems but it should be possible to establish the U^I. characteristic
so definitely that changes of the plate current of more than a few percent
no longer occur under othen/^ise equal conditions for individual tubes
operating with a control range of about 1 V at the control grid« If the
grid current remains below 5 3C 10 A, it can then be different in
The great importance inherent in electrometer bridge systems leads
to a further urgent demand. For several decades, the requirements for
the production of electrometer twin-tetrodes have teen theoretically-
known. The point here is that both systems must be and remain very
similar to each other in operation. This makes it necessary that both
control grids in the twin-tube are constructed with very high-degree
insulation. An effort should be made that both anodes receive electrons
/?/ /German has "Elektroden" = "Electrodes"— Tr._/ from the same parts of
the cathode fl, 10? 15/» Moreover, the space-charge grids of both
systems should be constructed separately. Indirect heating would somewhat
simplify the stabilization problems. If satisfactory twin- tubes do not
exist, it would be highly desirable to be able to order selected pairs of
tubes for the bridge system which was generally possible in the past.
e) Industrial Equipment using Electrometer Tubes . Stange & ¥olfr^^m,
Berlin S¥-6l, produces apparatus for continuous recording of the four
atmosph.eric-eleo1;ric basic elemen"ts ("atmospheric-electric station")
which includes four electrometer-tube bridge amplifiers for operation by
connection to a network power source* The high-ohmic design of the grid
at the compensation aggregates of the individual bridges permits compen~
sation of disturbing voltages from the atmospheric-electric antennas and
the measurement circuits,
"Teraohmmeter" utilizing electrometer-tube bridge systems are
produced by the company E. Jahre, Berlin ?/-35« They permit measurement
of very high resistances even at low voltages which is not the case for
most of the other teraohmmeters,
Keithley Instruments, Inc., Cleveland (Ohio, USl) furnishes a whole
line of equipment provided with electrometer tubes which is suitable for
tube-electrometers, tube-galvanometers, teraohmmeters, etc. for many
purposes. Types 510> 610, 610-A, 600tA, ..411»-412, 413," 410, 420 contain
two electrometer tubes in a bridge circuit (but compensation side is not
high-ohmic) J types 200, 200A, 414> etc., contain only one electrometer
tube (this listing is not complete).
Yictoreen Instrument Co., Cleveland, Ohio, UBA, a producer of widely
employed electrometer tubes, furnishes equipment provided with electro-
meter tubes, including types YTE-p to 7TE-3.
The electrometer of the company P. 1. Klein, Tettnang/Bodensee, has
an electrometer tube as input tube.
Literature references for the part II were published in the preceding
issue (J 8334-3, December I96I).
■P Mv |]
'H Ti a
tj rH -H
i rfi Eior '
< U oocd
( © tom
flJO cdOi o
W -rs feci
si H (3
rS O r* S O
1 i I ■' ■ - ■
C4 t-t r-< pH C4
j 833^ -A
o -o «
I c^ «>- c*«
s:1 top4 i
^ © '
© x! .. .
(HI O O
, o I
o ^^ o
03 p( P^
O 02 ««^
t^ *r-j r^
r-i -p e
fQ II II
rt 9 y
rt « « * ■* •*
C. 00 ■ CO
3 c5 -^ Q
s o c
O r-* .-« c*. O N
■ O ■^ iM C d
»-* O- o
O O O ift o
ift W. O CO o
Q ca' d
t> -o "S '5 o
-o o o o ^
.2 -is "S £ .2
■ "u o S S *S
- r* r* P* ;i^ E-t
pH C-l ' .S .
.jft o r* 05
O O O ro
o o o o
CO 00 r1
00 ■«*•" 1>"
"- 1 ..-1 r>
t/i n -* -j3
= 2.5 c«